Illustration for: Klarent Raises $8M To Make Software Testing Deterministic

Klarent Raises $8M To Make Software Testing Deterministic

Klarent, a Zurich startup founded by two ex-Google engineers, raised an $8M seed led by Mosaic Ventures to sell AI agents that write and reuse test code instead of calling a model at runtime.

By the Numbers

$8M Seed
Round
Mosaic Ventures
Lead
25
Team size
~3,500
Daily tests, 1 customer
ShareXLinkedInEmail

THE RUNDOWN

1

Klarent's pitch is 'deterministic autonomy' -- AI writes and reuses test code instead of calling a model on every run, cutting cost and flakiness.

2

One customer, JD Sports, already runs about 3,500 tests a day on the platform across its order-to-delivery pipeline.

3

Founders Asheem Panakkat and Momchil Ivanov pivoted from building LLM-evaluation tools after deciding testing AI-generated code was the bigger opportunity.

4

QA is a roughly $35B market still mostly done manually, per Panakkat -- the real competitor is outsourced headcount, not another startup.

The VC Read

Value Add VC analysis

The founders already pivoted once, from LLM evals to QA automation, which usually means they're chasing a market rather than being chased by it -- worth checking how much of the 3,500-tests-a-day number comes from one anchor account versus genuine multi-customer traction before assuming the model generalizes. Mosaic's partner framed the bet as conviction in "determinism," which really means betting enterprises will pay a premium to avoid the cost and flakiness of agentic testing tools that call a model on every single run.

Analysis

Klarent raised an $8 million seed round led by Mosaic Ventures, with Moonfire Ventures also participating alongside angel investors including Harris Barton, AngelInvest Ventures and ex-Google director Jürgen Galler, according to Tech Funding News. The Zurich-based company automates enterprise software quality assurance.

Founders Asheem Panakkat and Momchil Ivanov met at Google, where Panakkat worked as a software engineer and engineering manager on Maps, Search, Shopping and Lens. The pair left in 2023 to start Klarent, first building evaluation tools for large language models before concluding that testing AI-generated code was the bigger opportunity and pivoting the company around it.

Klarent calls its approach "deterministic autonomy": AI agents write and reuse test code rather than making a fresh model call every time the software runs, with a human-in-the-loop service checking each result and a model still involved in writing and repairing tests. One customer, JD Sports, runs about 3,500 tests a day on the platform across processes spanning website orders through delivery systems.

“One customer, JD Sports, runs about 3,500 tests a day on the platform across processes spanning website orders through delivery systems.”

The competitive field is already crowded:

  • Octomind -- raised $4.8M in seed funding; Panakkat argues it can't run Klarent's approach at scale.
  • Tricentis -- bolting an agentic feature onto its existing product rather than rebuilding around determinism.
  • QA Wolf and Momentic -- both have raised larger US rounds competing in agentic QA.
  • LambdaTest -- raised $45M, a bigger and more established player in the category.
  • Traditional outsourcing -- the real incumbent, according to Mosaic partner Chandar Lal, who called it a "low-cost, brute-force" business worth hundreds of billions in revenue.

Panakkat put the opportunity in blunt terms: "Quality assurance is a $35 billion business, give or take, and most of it is carried out manually." Klarent's 25-person team will put most of the new money toward hiring and sales expansion in the US and UK, with engineering a secondary priority.

ShareXLinkedInEmail

Key Sources

2 sources

THE WIRE in your inbox— Tech, startup & VC news with The VC Read, a few times a week. Free to subscribe, no spam.